A rack-mountable server (hereafter referred to as a “server”) is typically housed in a server rack. A server rack is a device that holds a large number of servers, which are often identical or similar in design. While these servers may be on-site, often the servers are located at remote locations to save space or maintenance costs and to provide a physically secure location for the servers. One common approach is to install the servers in a server farm. A server farm typically includes anywhere from tens to multiple thousands of servers located together and used by one or more companies.
Regardless of the physical location of a server, a user generally interfaces with the server through a remote interface. In a typical implementation, a user interfaces with a server through a remote computer system connected to the server over a network. The server is typically accessed through a unique network address, which allows the user to connect to the server from the remote location.
If a server or a component within the server malfunctions or fails, it may become necessary to service the faulty server or component within the server. Security restrictions due to multiple companies sharing space in a server farm, a server farm being located in a remote location, or any number of other factors may prevent the user from personally servicing the faulty server or component within the server. To further complicate the service required for servers, some servers hold one or more field replaceable units (FRUs) such as removable hard drives or other components that may require frequent service.
As the server is typically accessed from a remote location via a network, in order to service a faulty server housed at a server farm, some identification associated with the faulty server is necessary to physically locate it among any number (potentially thousands) of identical servers.
Two conventional techniques have been employed in server farms to allow a service technician to locate a particular server in the server farm. If the server requiring service is still functioning on some level, then the remote user may trigger lights on the server to blink. Further, depending on the particular problem with the server, the remote user may trigger the lights to blink in certain manner (i.e., at a certain speed, for a specific duration, in a particular pattern, lights of only a specific color may blink, etc.). As an alternative to blinking lights, some servers may also emit a noise (such as a beep) to indicate that the server is faulty.
While the aforementioned techniques enable a service technician to locate a server in the server farm, servers that are entirely nonfunctional clearly cannot use either of the aforementioned techniques, as the aforementioned techniques require power. In such situations, the service technician may locate the particular server using the serial number of the server along with information about the possible location of the server in the server farm. Regardless of which technique is used to locate the server, locating a particular server to service may take a significant amount of time.